Abstract
As product-development cycles become shorter and companies demand more rapid achievement of reliability goals, it is becoming more and more important to use quantitative accelerated life tests (ALT) to predict and improve reliability. Today there is an abundance of methods to plan and analyze accelerated tests, but there are also many pitfalls. This paper identifies some major problems and concerns in conducting and interpreting the results of accelerated tests. We identify various pitfalls, such as using an equal unit allocation at all levels of an accelerating variable, using unnecessarily complex testing and data analysis schemes, attempting to predict life from a HALT, using ALT at the system level, not having adequate time-to-failure information, using extreme extrapolation, ignoring the impact of idle time in use-rate acceleration tests, and not understanding interactions between accelerating variables.
Additional information
Notes on contributors
William Q. Meeker
Dr. Meeker is Distinguished Professor of Liberal Arts and Sciences and Professor of Statistics at Iowa State University. He is a Fellow of the ASQ. His email address is [email protected].
Georgios Sarakakis
Mr. Sarakakis is Staff Reliability Engineer at Tesla Motors. His email address is [email protected].
Athanasios Gerokostopoulos
Mr. Gerokostopoulos is Reliability Engineer at ReliaSoft Corporation. He is a Certified Reliability Professional (CRP) and an ASQ Certified Reliability Engineer and Certified Quality Engineer. His email address is [email protected].